Lesson 2: How do engineers find out what’s underground?

You will learn

Seeing the ground you can’t see

You can’t see the ground you’re building on. Ground investigation is how engineers find out what’s there, how strong it is and where the water is, before design decisions are fixed and before surprises become expensive.

You will need a calculator for a few short calculations.

Why it matters

Why investigate the ground at all?

Soils and rocks form through irregular natural processes: weathering, transport and deposition. As a result, the ground can change completely over a few metres, both across a site and with depth. Two boreholes 20 m apart can tell very different stories.

Without reliable information about the ground, a designer has two bad options:

• Over-design: assume the worst everywhere. The structure is safe, but costs more and uses more material than it needs to.
• Under-design: assume the ground is better than it is. This risks excessive settlement, damage or failure, and in the worst cases, loss of life.

Ground investigation costs money and time, and because the ground is out of sight, clients are often tempted to cut it back. Yet unexpected ground conditions are one of the most common causes of delays and cost overruns in construction. A well-planned investigation is usually a small fraction of the cost of the problems it prevents (Site Investigation Steering Group, 1993).

How much investigation a project needs depends on:

• the type, size and sensitivity of the structure, and the consequences if it goes wrong
• how complex and variable the geology is, including its history (for example old mining, landslides or filled ground)
• how much is already known about the area
• codes of practice, regulations and the client’s requirements and budget.

The big picture

The stages of a ground investigation

A good investigation is planned and carried out in stages. Each stage uses what the previous one found, so effort and money go where they’re most useful (Clayton, Matthews and Simons, 1995; BSI, 2020).

Stages 1 and 2 are cheap and quick. Stages 3 and 4 (the fieldwork and laboratory testing) are usually the most expensive, because they need specialist rigs, crews, equipment and time. That’s exactly why stages 1 and 2 matter: they make sure every expensive borehole is in the right place, to the right depth, with the right tests.

Stage 1

Desk study: what you can learn before visiting

The desk study gathers everything that is already known about the site and its surroundings. It’s carried out in the office, before anyone visits.

Much of this is now held in digital maps and databases (geographical information systems, or GIS). In the UK, for example, the British Geological Survey publishes geological maps and many borehole records online.

Stage 2

Walkover survey: reading the ground

The walkover survey (also called a reconnaissance) is the first visit to look at the ground. You take the desk study findings with you, check them against what’s there today, and look for clues about the ground. Record everything with notes, sketches and photographs:

• Landform: slopes, hollows, mounds, cuttings and embankments, and signs of past ground movement such as cracked or bulging ground, tilted trees or fences, or uneven surfaces that may indicate landslips.
• Exposed soil and rock: in banks, cuttings, quarries, stream beds or recent excavations nearby.
• Signs of made ground: unusually flat or raised areas, rubble, ash, brick or other debris, and stained or discoloured ground, which can also point to contamination.
• Water: springs, seepage, wet or boggy ground, reeds and other water-loving plants, ponds and drainage.
• Vegetation: especially large trees near clay soils, which can shrink clay as they draw water from it.
• Buried services: covers, marker posts and other signs of services that must be located before any digging or drilling.
• Ground conditions that affect the fieldwork: soft ground, steep slopes and overhead cables, which limit where investigation equipment can safely work.
• Existing investigation points: old boreholes, pits or trenches.
• Existing buildings on or near the site, and how they have performed.

Stage 3: the plan

Planning the fieldwork: how many holes, and how deep?

The fieldwork (often called soil exploration) has four main objectives:

1. The ground profile: the sequence, thickness and lateral extent of each soil and rock layer, and the depth to bedrock where relevant.
2. Representative samples: to identify and classify the soils and rocks, and to test them in the laboratory.
3. Groundwater conditions.
4. In-situ tests: to measure properties of the ground where it lies.

A fieldwork plan states the methods to use, the number and positions of investigation points, the sampling and testing required, and how groundwater will be observed.

How far apart?

Eurocode 7 Part 2 gives indicative spacings for investigation points (BSI, 2007). They are a starting point, not a rule. Points are placed closer where the ground is variable or the structure is sensitive, and further apart where the ground is uniform and well known.

How deep?

The investigation must go deep enough to cover all the ground that the structure will stress, plus any weak layers below that could still affect it. As a starting point, Eurocode 7 Part 2 suggests that for high-rise structures and civil engineering projects on pad or strip foundations, the depth investigated below foundation level should be the larger of 6 m and 3 times the foundation width. For rafts, it suggests at least 1.5 times the raft width (BSI, 2007). Investigations can stop shallower where strong rock is proven, and must go deeper where weak layers or piles are likely.

Stage 3: the methods

Pits and boreholes

Exploration methods fall into two groups:

• Intrusive (destructive) methods dig or drill into the ground. You can see and sample the soil directly, so they’re the most reliable and most widely used.
• Non-intrusive (non-destructive) methods, mainly geophysics, measure the ground from the surface without disturbing it. They are covered at the end of this page.

The choice depends on the ground conditions, the depth needed, access, the information required, time and budget. Most investigations combine several methods.

Trial pits

A trial pit is a hole dug with a mechanical excavator, usually to about 4–5 m, depending on the machine’s reach and the ground. It is simple, quick, relatively cheap and very reliable for shallow ground, because you see a whole face of soil, including layering, fissures, fill and obstructions.

• Large disturbed samples come from the spoil. High-quality block samples can be cut by hand from the side or base of a safely supported pit.
• Pit sides can collapse without warning. Never enter an unsupported pit. UK practice generally prohibits entry into unsupported excavations deeper than 1.2 m, so most logging is done from the surface.
• Below the water table, pits can flood and sides collapse, especially in sands, so they may need dewatering or may not be possible.
• Pits leave a large area of disturbed ground, so they must be backfilled and compacted, and should not be dug where a foundation will go.

Window (dynamic) sampling

Steel tubes, usually about 1 m long and lined with a plastic sleeve, are driven into the ground by a percussive hammer and pulled out full of soil. Small, often tracked rigs can work in tight spaces such as gardens and inside buildings. Depths are commonly around 5–10 m, depending on the ground. It’s fast and good for shallow investigations, but the samples are disturbed and the method struggles in dense gravels and stiff, stony ground.

Cable percussion boring

This is the standard way to drill through soil in the UK. A tripod rig repeatedly lifts and drops a heavy tool on a cable (typically over 1–2 m) to advance a borehole, usually 150–300 mm in diameter. A heavy sinker bar may be added above the tool to increase the impact. The tool depends on the soil (Knappett and Craig, 2019):

• Shell (bailer): for sands and gravels. A heavy tube with a flap valve at the bottom that keeps loose soil inside.
• Clay cutter: for clays and silts. An open tube with a cutting shoe that grips cohesive soil.
• Chisel: breaks up cobbles, boulders and hard layers.

Steel casing is driven down behind the borehole to support its sides. Boreholes can reach several tens of metres. The impact disturbs the soil at the base of the hole, so the base is cleaned out before samples are taken or tests are run.

Rotary drilling

A rotating drill string advances the hole, with water, air, mud or foam (the flush) cooling the bit and carrying cuttings to the surface.

• Rotary open-hole drilling grinds the ground into cuttings. It’s fast but gives no proper samples, so it’s used to get through ground quickly or to check for voids.
• Rotary coring uses a hollow bit and a core barrel to recover a solid cylinder (core) of rock or very stiff soil. It is the main way to investigate rock and to prove rockhead: a boulder can look like bedrock, so the core must continue for several metres into rock to confirm it (commonly at least 3 m in practice).

Core quality is described with measures such as the Rock Quality Designation (RQD): the total length of intact core pieces at least 100 mm long, as a percentage of the length drilled (Deere, 1964). A high RQD means few fractures.

Augers

Augers are screw-shaped or bucket-shaped tools turned into the ground by hand or by machine:

• Hand augers: small diameter and shallow, useful for quick checks of soft soils.
• Flight augers: a continuous spiral that lifts soil to the surface. Hollow-stem augers let samplers pass down through the middle while the auger supports the hole.
• Bucket augers: large-diameter holes in suitable soils.

Augering works best where the hole stands up without support, above the water table. Casing is awkward because the auger must be removed to install it, although bentonite slurry can support unstable holes.

Wash boring

Water is pumped down hollow rods and jets out at a chisel bit, loosening the soil, which is washed back up and collected. Changes in the feel of the rods or the colour of the returning water show a change of layer. Wash boring is quick in many soils but slows in coarse gravel and larger particles. The washed-out material is not acceptable as a sample, but the method is useful for advancing a hole to a depth where a proper sample can be taken. It is used more in some other countries than in the UK.

Without digging

Geophysics: looking without digging

Geophysical methods measure contrasts in physical properties of the ground from the surface (or between boreholes): how fast waves travel, how easily electricity flows, how radar signals reflect, or tiny changes in gravity or magnetism. Because different soils, rocks, water and voids have different properties, the measurements can show where boundaries and anomalies are (Clayton, Matthews and Simons, 1995).

Strengths: quick, relatively cheap per area covered, no damage to the site, and good for filling gaps between widely spaced boreholes or showing where extra boreholes are needed. It’s also useful on sensitive or contaminated sites.

Limitations: results are interpretations that can have more than one explanation. Geophysics must be designed and interpreted by experienced specialists, and checked against boreholes or pits (“ground truthing”). It supports intrusive investigation; it doesn’t replace it.

What comes out of the hole

Samples: disturbed and undisturbed

Samples let engineers identify the soil and test it in the laboratory. What you can test depends on how much the sample was disturbed:

• Disturbed samples keep the soil’s particles and (if sealed) its water content, but not its structure. They’re fine for identification, particle size and plasticity tests.
• Undisturbed samples keep the soil’s structure and density as nearly as possible. They’re needed for tests of strength, stiffness, compressibility and permeability.

No sample is perfectly undisturbed: removing soil from the ground releases the stresses that were acting on it. Eurocode 7 Part 2 ranks samples in quality classes from 1 (highest, suitable for strength and compressibility tests) to 5 (only shows the sequence of layers) (BSI, 2007).

Common samplers

• Bulk and small disturbed samples: soil placed in bags, tubs or jars, from pits, window samples or the borehole tools.
• Split-spoon sampler: a thick tube that splits lengthways, used in the Standard Penetration Test. Gives a disturbed sample.
• Open-drive tube samplers: tubes driven or pushed into the base of a borehole. Thin-walled tubes pushed in steadily (rather than hammered) give much better samples of clay than thick-walled tubes.
• Piston samplers: a piston inside the tube holds the sample in place as the tube is pushed and withdrawn, giving high-quality samples of soft clays. They are more costly.
• Block samples: cut by hand from a pit, usually the highest quality possible, but only at shallow depth.

How thick is too thick? The area ratio

The thicker a tube’s wall, the more soil it pushes aside and the more it disturbs the sample. This is measured by the area ratio (Hvorslev, 1949):

The lower the area ratio, the less disturbance. Well-designed thin-walled samplers have area ratios of around 10% or less.

Labels you’ll see on logs

UK borehole logs commonly use letter codes for sample types, for example B (bulk disturbed), D (small disturbed), U or UT (undisturbed tube, with UT for thin-walled), W (water) and ES (environmental soil, for chemical testing). Always check the key on the log, because codes can vary between companies.

Testing in the ground

In-situ tests

Some properties are best measured in the ground itself, at the real stresses and water conditions, especially in soils that can’t be sampled undisturbed. The most widely used in-situ tests are (Clayton, Matthews and Simons, 1995; Knappett and Craig, 2019):

The water

Groundwater

Groundwater affects almost every part of geotechnical design and construction. It changes how strong the soil is, decides whether excavations will stay stable or need pumping (dewatering), can push buried structures upwards (uplift), can attack concrete, and decides which construction methods and timings are possible. Missing it is one of the most common causes of problems on site.

Groundwater conditions you may find

• Unconfined water table: the level below which the ground is saturated, free to rise and fall.
• Perched water: a local body of water held up by a less permeable layer, such as a clay lens, above the main water table.
• Confined (artesian) water: water in a permeable layer trapped under a low-permeability layer. It’s under pressure, so when a borehole reaches it, the water rises up the hole, sometimes above ground level.

How groundwater is measured

• During drilling: the driller records the depth of every water strike, then watches whether the level rises, typically recording it after 20 minutes, and notes when casing seals the water off.
• Over time: a standpipe or standpipe piezometer is installed in the borehole, with a slotted section (the response zone) sealed into the layer of interest. Readings are taken over weeks or months. Electronic piezometers can log pore water pressure continuously.

Safety first

Contaminated ground and safety

If the desk study shows that the ground may be contaminated, for example from former industry, landfill, fuel storage or made ground, the investigation must be extended. Contaminants may be solids, liquids or gases, and can include harmful chemicals, asbestos fibres, toxic or explosive gases, biological agents or radioactive material. Some contaminants stick to the surfaces of fine soil particles.

Contamination affects the whole investigation and may change the foundation design or even the use of the site. The investigation must (BSI, 2017):

• protect the health of everyone on site and everyone handling samples
• prevent contamination spreading, whether carried by people and equipment, by surface water or groundwater, or by wind
• take samples for chemical testing using the right containers and methods.

Every investigation, contaminated or not, also needs a safety plan covering buried and overhead services, unexploded ordnance where relevant, unstable excavations and contaminated ground. Before breaking ground, service plans are checked, the area is scanned with a cable locator, and a hand-dug inspection pit is commonly excavated at each drilling position (Health and Safety Executive, 2014).

Stages 4 and 5

From fieldwork to report

Laboratory testing

Samples go to a laboratory, where tests are chosen to answer the design questions (Knappett and Craig, 2019):

Borehole and trial pit logs

The findings from each hole are summarised on a log, one per hole. Layouts vary between companies, but a log normally shows:

• project, hole number, location coordinates, ground level, dates and method of drilling
• the depth, level and description of each layer (stratum), with a graphic legend
• samples and in-situ tests, with their depths and results (for example SPT N values)
• water strikes and levels, casing depths, and any installed standpipes
• remarks, such as obstructions or why the hole was stopped.

In the UK, levels are usually given in metres above Ordnance Datum (m OD), the national reference level. Other countries use their own national datum. Depth and level are linked by:

The reports

• Factual report: what was done and found, including methods, logs, test results, groundwater readings and plans, with no design opinions.
• Interpretative report: what the facts mean, including the ground model, design parameters, hazards and recommendations.

Under Eurocode 7, the factual results and their evaluation are presented together in a Ground Investigation Report, and the geotechnical design itself is recorded in a Geotechnical Design Report (BSI, 2007).

A good report clearly describes the soils (types, state and layering), the methods used, all test methods and results, groundwater conditions, and anything unusual, such as unexpected soil types, frost-susceptible soils or waterlogged areas. Maps, sections, photographs and diagrams help to communicate the findings.

Worked examples

Putting it together

Your turn

Five questions to try

Quick check

Five quick questions

Choose an answer to see the explanation.

Key points

What to take away

• The ground varies across a site and with depth, so investigation protects against both over-design and under-design.
• Ground investigation runs in stages: desk study → walkover → fieldwork → laboratory testing → reporting, with more investigation during design and construction if needed.
• Plan the number, spacing and depth of points from the structure, the geology and published guidance, then apply judgement.
• Choose methods to suit the ground, depth, access and the question: pits, window sampling, cable percussion, rotary coring, augers, geophysics.
• Undisturbed samples are needed for strength, stiffness, compressibility and permeability tests. Lower area ratios mean less disturbance.
• In-situ tests (SPT, CPT, vane, pressuremeter, dilatometer) measure the ground where it lies. Correlations give first estimates only.
• Groundwater must be monitored over time, not just noted during drilling.
• Safety comes first: trace services before digging, and treat possible contamination seriously.

Assessment (optional)

Test yourself properly

Unlike the practice questions, this assessment has no answers on the page. Attempt it on your own, as if it were an exam.

Section A: Concepts (12 marks)

1. A client says a ground investigation “just means drilling a few boreholes”. Explain why this is not correct. [2 marks]
2. Name four sources of information you would use in a desk study and state what each one would tell you about a site. [4 marks]
3. Explain the difference between a disturbed and an undisturbed sample. Give one laboratory test that can be done on each, and explain why the undisturbed sample is needed for your second test. [4 marks]
4. Give one reason a client might want to reduce the amount of ground investigation, and describe one possible consequence of doing so. [2 marks]

Section B: Calculations and interpretation (16 marks)

5. A simplified log for borehole BH3 (ground level 31.80 m OD):

   Depth (m)	Description	Tests and water
   0.00–0.30	Topsoil
   0.30–2.60	Soft grey silty CLAY
   2.60–6.40	Loose to medium dense brown SAND	Water struck at 2.90 m, rose to 2.00 m after 20 min. SPT at 4.0 m: N = 8
   6.40–12.00	Stiff brown CLAY	SPT at 9.0 m: N = 22

   (a) Calculate the levels (m OD) of the top and the base of the sand. [2 marks]

   (b) Describe what the water observations suggest, and explain how you would establish the long-term groundwater level. [3 marks]

   (c) Using c_u ≈ f₁ N with f₁ = 5 kPa, estimate the undrained shear strength of the clay at 9.0 m, and state one limitation of this estimate. [3 marks]

6. Two tube samplers are available. Sampler P has outside and inside diameters of 102 mm and 98 mm. Sampler Q has outside and inside diameters of 114 mm and 100 mm.

   (a) Calculate the area ratio of each sampler. [2 marks]

   (b) Which sampler would you choose to take a sample of soft clay for an oedometer test, and why? [2 marks]

   (c) Explain why even a high-quality tube sample of clay is never perfectly undisturbed. [2 marks]

   (d) Explain why undisturbed sampling of clean sand is difficult, and name an in-situ test you could use instead to assess its density. [2 marks]

Section C: Applied case (12 marks)

7. A two-storey extension (40 m × 25 m) is planned at a busy primary school, on strip footings about 0.8 m wide at 1.0 m depth. Historical maps show the area was a brick pit in the 1900s that was later backfilled. The existing school building next to the site has diagonal cracks above some windows. Mature oak trees line the site boundary, and the desk study shows clay beneath any fill.

   Write a short ground investigation proposal (maximum 400 words) that covers:

   • what you would add to the desk study, and why [2 marks]
   • what the walkover observations might mean, including the cracks and the trees [2 marks]
   • the exploration methods and an outline layout, with reasons [3 marks]
   • the samples, in-situ tests, laboratory tests and groundwater or gas monitoring you would include [2 marks]
   • the key safety considerations for this site [1 mark]

   Clear, well-organised communication suitable for a client. [2 marks]

How your work is judged

References

• British Standards Institution (BSI) (2007) BS EN 1997-2:2007 Eurocode 7. Geotechnical design. Ground investigation and testing. London: BSI.
• British Standards Institution (BSI) (2017) BS 10175:2011+A2:2017 Investigation of potentially contaminated sites. Code of practice. London: BSI.
• British Standards Institution (BSI) (2020) BS 5930:2015+A1:2020 Code of practice for ground investigations. London: BSI.
• Clayton, C.R.I., Matthews, M.C. and Simons, N.E. (1995) Site Investigation. 2nd edn. Oxford: Blackwell Science.
• Deere, D.U. (1964) ‘Technical description of rock cores for engineering purposes’, Rock Mechanics and Engineering Geology, 1(1), pp. 17–22.
• Health and Safety Executive (2014) Avoiding Danger from Underground Services. HSG47. 3rd edn. London: HSE Books.
• Hvorslev, M.J. (1949) Subsurface Exploration and Sampling of Soils for Civil Engineering Purposes. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station.
• Knappett, J.A. and Craig, R.F. (2019) Craig’s Soil Mechanics. 9th edn. Boca Raton, FL: CRC Press.
• Site Investigation Steering Group (1993) Site Investigation in Construction 1: Without Site Investigation Ground is a Hazard. London: Thomas Telford.
• Stroud, M.A. (1974) ‘The standard penetration test in insensitive clays and soft rocks’, Proceedings of the European Symposium on Penetration Testing, Stockholm, 2(2), pp. 367–375.

Further learning

• Read: Clayton, Matthews and Simons’ Site Investigation is a thorough, practical guide to every stage covered in this lesson. Many libraries hold a copy.
• Explore real borehole records: in the UK, the British Geological Survey’s free GeoIndex map (bgs.ac.uk) lets you view thousands of real borehole logs. Many national geological surveys publish similar records. Find one near you and practise reading it.
• Safety around services: the HSE guide Avoiding Danger from Underground Services (HSG47) is free to download from hse.gov.uk.
• Industry guidance: the Association of Geotechnical and Geoenvironmental Specialists (ags.org.uk) publishes free guidance and the AGS data format used to exchange borehole data.
• See it done: look for videos of cable percussion rigs, window sampling and cone penetration testing to see how each method works on site.